There are just four days left before the official release of Solving for Project Risk Management. In anticipation of this event I have been blogging about individual book chapters. Previously this week I discussed chapters that focus on cost growth and schedule delays, the need for quantitative analysis of risk, and the tendency to underestimate risk. In today’s post, I discuss Chapter 5, which is titled “The Portfolio Effect and the Free Lunch.” The idea of the free lunch dates back to nineteenth century saloons in America, which would offer a free lunch. However, the food would be oversalted, leading to customers buying more drinks as a result. The so-called “free lunch” usually led to customers paying more than they would have otherwise.
One of the first mentions of the free lunch in an economic context appeared in a short piece printed in the El Paso Herald Post in 1938 titled “Economics in Eight Words.” This is a fable in which a king asks his advisers to summarize economics in a brief and simple manner. They respond with 87 volumes each 600 pages long, leading the king to order the execution of half of his advisers. Further demands led to shorter summaries, none of which satisfied the king, and more executions took place. Finally, the last economist, faced with loaded crossbows aimed at his forehead said, “Sire, in eight words I will reveal to you all the wisdom that I have distilled through all these years from all the writings of all the economists who once practiced their science in your kingdom. Here is my text: ‘There ain’t no such thing as free lunch.'”
In modern day economics, diversification has been called “The only free lunch on Wall Street.” Rather than do full-up portfolio analysis, organizations often conduct risk analysis only at the project level, relying on the purported benefits of diversification, termed a “portfolio effect.” For example, if 10 projects are funded at the 70th percentile, the total should be higher than the 70th percentile for the total organization. However, the benefits of diversification depend upon the relationship between risk and reward. When risk is assessed realistically, the benefits are quite small. Moreover, many projects attempt to fund to a low percentile of the cost risk distribution in the hopes of still achieving a high percentile for the entire organization. If the funding is below the mean, however, a negative portfolio effect is guaranteed, which means that the organization is riskier than any individual project. I have seen this in practice. An organization where I worked funded major projects to the 50th percentile (because risk is skewed, the 50th percentile is usually below the mean). When I analyzed the risk of the portfolio, I found the overall percentile for the portfolio of these projects was much lower. Not having enough funding at the project level is bad, but funding projects so that there is no reserve at the organization level is a recipe for disaster. As an example, consider the 10 project in the table below. Each is funded at the 50th percentile, but the total portfolio is only funded at the 36 percentile. An additional 18% is needed just to fund the portfolio to the 50th percentile!
Funding should not be set to low levels in the hope that a mythical portfolio effect will result in a high percentile for the entire organization. Projects should never be funded below the mean – there should be some levels of reserves held to achieve at least the mean. Also, there is no shortcut to achieving high confidence levels. In an informal survey I conducted during a presentation to over 100 people, more than 80% responded that their organizations do not conduct portfolio-level risk analysis. As we have seen, there is no shortcut to getting high confidence levels for an entire organization. Everyone should be conducting portfolio-level risk analysis. There ain’t no such thing as a free lunch, and there ain’t no such thing as a portfolio effect either.